Meta-analysis of the benefit of beta-blockers for the reduction of isolated atrial fibrillation incidence after cardiac surgery

Figures

• 

  Forest plot of isolated POAF incidence after cardiac surgery.

• 

  Figure 1. PRISMA flow diagram of literature retrieval. A framework that describes the process of our study selection—519 papers were identified from databases and other sources, 414 studies were screened after duplicate removal, final 21 articles assessed in full text, and 17 RCTs were eventually included in our study.

• 

  Figure 2. Forest plot of isolated POAF incidence after cardiac surgery. A forest plot comparing isolated POAF incidence between b-blocker and control users, in our 17 included trials. Overall risk ratio of 0.52 (95% confidence interval, 0.41-0.66; P = .31) suggests a 48% reduction in risk of isolated POAF in b-blocker users among our 17 included trials. Size of the blue square represents the relative weight of the studies' contributions to the overall risk ratio. MH, Mantel–Haenszel; CI, confidence interval.

• 

  Figure 3. Subgroup analysis on the influence of type of surgery. The forest plot suggests an overall reduction in isolated POAF risk for CABG and valve surgeries, where ONCABG displays the lowest risk ratio (risk ratio, 0.34 [0.04-3.15], P = .56, I2 = 0%). Size of the blue square represents the relative weight of the studies' contributions to the overall risk ratio. MH, Mantel–Haenszel; CI, confidence interval; CABG, coronary artery bypass grafting; ONCABG, on-pump coronary artery bypass grafting; OPCABG, off-pump coronary artery bypass grafting.

• 

  Figure 4. Subgroup analysis on the route of b-blocker administration. The forest plot indicates a reduction in isolated POAF risk in all types of b-blocker administration, although the IV route yields the lowest isolated POAF risk. (risk ratio, 0.49 [0.32-0.75], P = .10, I2 = 39%). Size of the blue square represents the relative weight of the studies' contributions to the overall risk ratio. MH, Mantel–Haenszel; CI, confidence interval; IV, Intravenous.

• 

  Figure 5. Subgroup analysis on the starting time of b-blocker administration. The forest plot displays an overall reduction in risk of isolated POAF incidence for all timings, but postoperative b-blocker administration is noted to have the lowest risk (risk ratio, 0.43 [0.29-0.62], P = .84, I2 = 0%). Size of the blue square represents the relative weight of the studies' contributions to the overall risk ratio. MH, Mantel–Haenszel; CI, confidence interval.

• 

  Figure 6. Subgroup analysis on the type of b-blocker. The forest plot shows a reduction in risk of isolated POAF incidence for atenolol, landiolol hydrochloride, metoprolol, and propranolol. In contrast, esmolol is found to have an increase in isolated POAF risk (risk ratio, 1.03 [0.36-2.92], P = .64, I2 = 0%). Size of the blue square represents the relative weight of the studies' contributions to the overall risk ratio. MH, Mantel–Haenszel; CI, confidence interval.

• 

  Figure 7. Subgroup analysis on dosage for IV landiolol hydrochloride. The forest plot depicts a reduction in isolated POAF risk for all dosages of IV landiolol hydrochloride, and the greatest reduction in risk is found from a dose of 2 μg/kg/min (risk ratio, 0.27 [0.20-0.36], P = .95, I2 = 0%). Size of the blue square represents the relative weight of the studies' contributions to the overall risk ratio. MH, Mantel–Haenszel; CI, confidence interval.

• 

  Figure 8. Funnel plot of publication bias. Our funnel plot of publication bias did not have any signs of asymmetry and hence, did not indicate any publication bias across our included studies.